JP6478643B2 - Information processing apparatus, information processing method, and program - Google Patents

Description

  The present invention relates to an information processing apparatus, an information processing method, and a program.

  In recent years, surveillance cameras have been widely used, and analysis techniques using flow lines (individual movement trajectories) have been developed as analysis techniques in surveillance camera systems. For example, in order to grasp the movement of a moving body in a monitoring space, a system that displays a flow line in real time is widespread. When displaying a flow line, it is required to display efficiently according to behavioral characteristics such as abnormal behavior to enable efficient grasping of the situation.

  In Patent Document 1, as a technique related to the display of the flow line, when the moving line is displayed by connecting the points where the moving body is observed with line segments, the flow line is displayed according to the time width between the observation points. A technique for changing the display form is disclosed. In this technique, a section in which the position information of the moving body cannot be observed is displayed in a different form from the other sections, thereby displaying a flow line indicating the actually moved path, an interpolated flow line, and the like. Can be distinguished. Patent Document 2 discloses a technique for changing whether or not to superimpose a flow line of a target moving body on a captured image in accordance with the moving direction of the moving body captured by an in-vehicle camera. In this technique, since the flow line of the target moving body is displayed only when approaching the current position of the vehicle or the predicted course, it is possible to efficiently grasp only the movement of the moving body that is likely to contact the vehicle.

  Further, in Patent Document 3, as a technique related to the cause display of abnormal behavior, it is estimated that an obstacle or the like causing the abnormal behavior of the moving body is present at the abnormality occurrence location, and the estimated abnormality occurrence location is displayed. A technique for helping to understand the cause of abnormal behavior is disclosed.

Japanese Patent No. 5176440 Japanese Patent No. 5484535 Japanese Patent No. 4118674

  There is a high possibility that a moving body that takes abnormal behavior is performing an unpredictable movement. Therefore, in order to present what abnormality has occurred, it is most effective to display the action trajectory itself represented by the flow line of each moving body. At this time, in the display of the flow line, if the entire length of the flow line of the moving body existing within the monitoring range is displayed, there is a problem that visibility is lowered and it is difficult to catch individual movements. However, when only the latest flow line is displayed, there is a problem that it is impossible to guess what kind of abnormality has occurred from the displayed flow line when there is an abnormal behavior that cannot be understood without observing for a long time.

  The present invention has been made in view of such problems, and displays the movement trajectory of the moving object so that the observer can easily distinguish the action to be observed from other actions among the actions of the moving object. Objective.

  Therefore, the present invention is an information processing apparatus, and based on the movement history of the moving object, the detection unit that is set as an observation target and detects the special action of the moving object, and the special action is not detected A moving trajectory of the mobile object is displayed with a first predetermined length, and when the special action is detected, the moving trajectory is longer than the first length, And display processing means for displaying in a length of 2.

  ADVANTAGE OF THE INVENTION According to this invention, the movement locus | trajectory of a moving body can be displayed so that an observer can identify easily the action which becomes an observation object among other actions.

It is a figure which shows the hardware constitutions of information processing apparatus. It is a figure which shows the software structure of information processing apparatus. It is a figure which shows the example of a display of a flow line display screen. It is a figure which shows an example of the data of a flow line. It is explanatory drawing of a shaping flow line data generation process. It is a figure which shows an example of a normal action model. It is a figure which shows an example of abnormality detection data. It is a figure which shows an example of the data for a display. It is a flowchart which shows a flow line display process. It is a flowchart which shows an abnormal action detection process. It is a flowchart which shows a display process. It is a figure which shows the structure of monitoring space. It is a figure which shows the example of a display of a flow line display screen. It is a figure which shows an example of an abnormal action model. It is a figure which shows symbol data. It is a figure which shows an example of the abnormality detection data which concern on 2nd Embodiment. It is a figure which shows an example of the data for a display which concerns on 2nd Embodiment. It is a flowchart which shows a display process. It is a figure which shows an example of a locus | trajectory display screen. It is a flowchart which shows a display process.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

(First embodiment)
The information processing apparatus according to the present embodiment displays a flow line that is a movement locus of a person based on an image of a person as a moving body that exists in a monitoring space (target space) that is a processing target. FIG. 1 is a diagram illustrating a hardware configuration of the information processing apparatus according to the first embodiment. The information processing apparatus includes a CPU 101, a ROM 102, a RAM 103, a storage unit 104, a display unit 105, an operation unit 106, and a communication unit 107. The CPU 101 controls the entire apparatus by executing a control program stored in the ROM 102. The RAM 103 temporarily stores various data from each component. The RAM 103 expands the program so that the CPU 101 can execute it. The storage unit 104 stores various data. As a medium of the storage unit 104, a flash memory, an HDD, a DVD-RAM, or the like can be used. The display unit 105 is composed of a liquid crystal panel or the like and displays various data. The operation unit 106 includes operation buttons, a touch panel, and the like, and receives instructions from the user. In addition, this apparatus can communicate with other apparatuses such as a monitoring camera via the communication unit 107. Note that the functions and processing of the information processing apparatus to be described later are realized by the CPU 101 reading a program stored in the ROM 102 or the like and executing this program. As another example, each unit illustrated in FIG. 2 may be realized by hardware such as an electronic circuit, and the functions and processing of the information processing apparatus may be realized by each hardware.

  FIG. 2 is a diagram illustrating a software configuration of the information processing apparatus. The information processing apparatus includes a monitoring video DB 201, a tracking unit 202, a data control unit 203, an accumulation flow line DB 204, a monitoring target flow line DB 205, a model construction unit 206, a behavior model DB 207, a detection unit 208, A display processing unit 209. The information processing apparatus receives a monitoring video imaged by an imaging device such as a monitoring camera and stores it in the monitoring video DB 201. The tracking unit 202 analyzes the monitoring video and extracts a flow line that is a movement locus of the person. Known techniques such as the background subtraction method can be used for detection of moving objects and extraction of flow lines. The extracted flow lines are stored in the accumulation flow line DB 204 and the monitoring target flow line DB 205 by the data control unit 203. The accumulation flow line DB 204 accumulates flow lines from the past to the present. The monitoring target flow line DB 205 stores only the flow line of the moving body existing in the monitoring space as the monitoring target at the latest time.

  The model construction unit 206 extracts a pattern from the flow line in the accumulated flow line DB 204 and constructs the behavior model DB 207. It is assumed that the model construction unit 206 performs analysis at a predetermined timing after a certain amount of flow line is accumulated in the accumulated flow line DB 204. The detection unit 208 performs real-time abnormal behavior detection processing using the constructed behavior model DB 207. The detection unit 208 compares the data stored in the monitoring target flow line DB 205 with the data stored in the behavior model DB 207, and detects abnormal behavior of the moving object in the monitoring space. Note that the detection unit 208 determines that a normal action has been detected except when an abnormal action is detected. The display processing unit 209 performs processing for displaying the flow line in the monitoring target flow line DB 205 on the display unit 105 according to the determination result of the detection unit 208. In the present embodiment, the display processing unit 209 displays a flow line from a certain past point to the reference point. Here, the reference point is a point in time when processing by the display processing unit 209 is executed, that is, a point in time when a flow line is displayed or a point just before being determined based on that point.

  FIG. 3 is a diagram illustrating a display example of the flow line display screen. FIG. 3A to FIG. 3D show transition of the flow line display screen according to the passage of time. On the flow line display screen, a symbol 301 indicating the position of each moving body at the display time point (reference point) and a flow line 302 as a moving locus of the moving body are displayed. In the present embodiment, the symbol 301 is indicated by a black circle, and the flow line 302 is indicated by a solid line. However, if the full length of the flow line of the moving body is displayed, the intersection of the flow lines increases and the visibility is lowered. Therefore, the information processing apparatus according to the present embodiment determines a past point at which a flow line is displayed, that is, a display start point, and displays only the flow line from the display start point to the reference point. The display start point is, for example, a time point 5 seconds before (the past) from the reference point.

  FIG. 3A is a diagram showing a flow line display screen at time “t = 0”. The symbol 301 indicates the position (reference position) of the moving object P at the reference point, and the flow line 302 indicates the moving locus from the display start point of the moving object P to the reference point, for example, the moving locus for the last 5 seconds. Yes. FIGS. 3B, 3C, and 3D are diagrams showing flow line display screens at the times 10 seconds, 20 seconds, and 30 seconds after FIG. 3A, respectively. 3A to 3D show a state in which the moving body P repeats reciprocation in the up-down direction and crawls in the space. It is assumed that such an ugly behavior is detected as an abnormal behavior at the time of FIG. In this case, the information processing apparatus according to the present embodiment changes the flow line 310 of the moving body P to the flow line 311 as illustrated in FIG. In other words, the information processing apparatus changes the length of the flow line of the moving body P to the length from the starting point of the abnormal behavior (time “t = 0”) to the reference point (time “t = 30”). .

  When abnormal behavior is detected, in order to improve the visibility of a monitor or the like, as shown in FIG. 3F, the line width of the flow line 320 of the moving body P in which abnormal behavior is detected is increased. It is also possible to highlight it. However, if the line width is simply increased, the observer can know that the moving body P has taken an abnormal action, but the last 5 seconds of movement of the moving body P cannot be distinguished from the normal action. I do n’t know that I ’m talking. On the other hand, as shown in FIG. 3E, the information processing apparatus according to the present embodiment displays a flow line from the starting point of the abnormal action. It is possible to confirm not only the occurrence of the occurrence but also the locus of abnormal behavior. Hereinafter, the flow line of the moving body in which the abnormal action is detected is referred to as an abnormal flow line, and the flow line of the mobile body in which the abnormal action is not detected is referred to as a normal flow line.

  FIG. 4A is a diagram illustrating an example of observation point data of flow lines obtained by the tracking unit 202. The observation point data includes a flow line ID, a time, a position, a speed, a partial space ID, and an acquisition source. The flow line ID is used to specify observation points by the same moving object, and a set of observation points (records) having the same flow line ID is handled as one flow line. The time is the observed time. The position and speed are the position and moving speed in the space at time, respectively. The partial space ID is an ID of a partial space where each observation point exists. By dividing the monitoring space into subspaces and using the IDs indicating the subspaces as position information of the observation points, the observation points in the same subspace can be regarded as existing at the same position. There is an effect of absorbing the difference in fine position. By controlling the division granularity of the partial space, it is possible to control how much position difference is allowed in the same pattern when the behavior model DB 207 is constructed. The acquisition source specifies an imaging device that images each observation point. When there are a plurality of imaging devices in the monitoring space, a plurality of imaging devices may be associated with one observation point.

  FIG. 4B is a diagram illustrating an example of shaping flow line data after preprocessing the data in FIG. The data control unit 203 performs conversion from the observation point data of the flow line to the shaped flow line data. The observation point data and the shaping flow line data are stored in the accumulation flow line DB 204 and the monitoring target flow line DB 205. The shaped flow line data includes a flow line ID and a partial space series. The subspace series represents the transition of the subspace corresponding to the observation point for each flow line ID as series data.

  FIG. 5 is an explanatory diagram of the shaping flow line data generation process. FIG. 5A is a diagram illustrating a state in which the entire monitoring space 500 is divided into partial spaces and a unique ID is assigned to each partial space. Here, the monitoring space 500 is divided at regular intervals, and a unique ID is assigned to each partial space, but may be divided using space structural knowledge (entrance / entrance / passage / section, etc.). Assume that a flow line 502 with a flow line ID “mov0001” as shown in FIG. 5B is detected. In this case, the data control unit 203 generates a subspace sequence “6, 3, 2, 3, 4” corresponding to the flow line 502 in FIG. 5B, and generates the flow line ID “mov0001”. To obtain the shaping flow line data of the flow line ID “mov0001”.

  FIG. 6 is a diagram illustrating an example of the abnormal behavior model stored in the behavior model DB 207. The behavior model DB 207 stores a flow line pattern as an abnormal behavior model. Here, the flow line pattern is a pattern extracted from a plurality of flow lines. For example, the model construction unit 206 constructs an abnormal behavior model in accordance with a user operation.

  FIG. 7 is a diagram illustrating an example of abnormality detection data generated by the detection unit 208. The abnormality detection data includes a flow line ID, a status, an abnormal action start point, and a detection point. Here, the status is information indicating the type of the detected action. In the present embodiment, the status takes a binary value of “normal” or “abnormal”. The start point is the time when the abnormal behavior is started. The detection point is the time when an abnormality is detected.

  FIG. 8 is a diagram illustrating an example of display data generated by the display processing unit 209. The display data includes a flow line ID, a display start point, and a display flow line. The display start point is information indicating which position of the total length of the flow line specified by the flow line ID is to be displayed. The display flow line is a coordinate value of a display flow line.

  FIG. 9 is a flowchart illustrating a flow line display process performed by the information processing apparatus. It is assumed that the extraction of the flow line and the construction of the behavior model DB 207 have been completed when the flow line display process is executed. The flow line display process performs a real-time process, and the information processing apparatus repeatedly executes the process periodically at predetermined time intervals. Then, the information processing apparatus ends the flow line display process when the software ends or when the apparatus pauses. In step S901, the detection unit 208 determines whether a predetermined time has elapsed since the previous flow execution. If the predetermined time has elapsed (Yes in S901), the detection unit 208 advances the process to S902. In step S <b> 902, the detection unit 208 acquires one shaped flow line data from among the shaped flow line data stored in the monitoring target flow line DB 205.

  Next, in S903, the detection unit 208 performs abnormal behavior detection processing based on the shaped flow line data. In step S904, the detection unit 208 determines whether or not the abnormal behavior detection process has been completed for all the flow lines stored in the monitoring target flow line DB 205. When the processing for all the flow lines is completed (Yes in S904), the detection unit 208 advances the processing to S905. If there is an unprocessed flow line (No in S904), the detection unit 208 advances the process to S902 and repeats the process. In step S905, the display processing unit 209 generates display data for all flow lines and displays a flow line display screen. Thereafter, the display processing unit 209 advances the processing to S901. As described above, by repeating the flow line display process, the detection unit 208 periodically detects abnormal behavior.

  FIG. 10 is a flowchart showing detailed processing in the abnormal behavior detection processing (S903) by the detection unit 208. First, in S1001, the detection unit 208 checks the status of the abnormality detection data of the flow line to be processed. When the previous abnormal behavior detection process (S903) is executed, the status of the abnormality detection data of the flow line to be processed is updated, so the detection unit 208 confirms the previous movement behavior detection process (S903). It becomes the hour status. If no abnormality detection data of the flow line to be processed has been generated, the detection unit 208 generates a record of abnormality detection data of the flow line to be processed. Specifically, the detection unit 208 records the flow line ID and the status in the abnormality detection data record. The detection unit 208 records “normal” in the status. On the other hand, the detection unit 208 does not record data in the abnormal action start point and the abnormality detection point of the abnormality detection data.

  If the status is “abnormal” (abnormal in step S1001), the detection unit 208 ends the abnormal behavior detection process because an abnormal behavior has already been detected. If the status is “normal” (normal in step S1001), the detection unit 208 advances the process to step S1002 to detect abnormal behavior. In other words, in S1001, the detection unit 208 identifies a moving object whose status is “normal” as a processing target, and excludes a moving object whose abnormality is already detected and whose status is “abnormal” from the processing target. Yes. That is, the process of S1001 is an example of a moving body specifying process. In step S1002, the detection unit 208 determines the presence / absence of abnormal behavior based on the shaping flow line data of the flow line to be processed. Specifically, the detection unit 208 determines whether or not abnormal behavior has been detected by comparing the abnormal behavior model illustrated in FIG. 6 with the shaped flow line data. Specifically, the detection unit 208 determines that the partial line sequence of the entire flow line to be processed up to the reference point is abnormal when it is included in the flow line pattern in the abnormal behavior model, and is not included. In this case, it is determined as normal. Here, the abnormal behavior is an example of a special behavior, and the process of S1002 is an example of a detection process for detecting the special behavior. Here, the abnormal behavior is determined according to the type of the moving body. In other words, the abnormal behavior model referred to by the detection unit 208 differs depending on the type of moving object to be observed.

  As another example, the detection unit 208 may perform abnormal behavior detection for a part of the flow line instead of performing abnormal behavior detection for the entire sequence length of the flow line. When abnormal behavior detection is performed on a partial flow line from a reference point up to a point in time, the past observation points in the flow line are excluded from the detection target. For this reason, the detection part 208 can determine again as normal, when it returns to normal action from abnormal action. In that case, the detection unit 208 refers to the abnormal action start point in the branching process of S1001. If a predetermined time has elapsed from the abnormal action start point, the process proceeds to S1002 regardless of the status. The abnormal behavior detection determination is performed again.

  When the abnormal behavior is not detected, that is, when the normal behavior is detected (No in S1002), the detection unit 208 advances the process to S1003. On the other hand, when an abnormal action is detected (Yes in S1002), the detection unit 208 advances the process to S1004. In step S1003, the detection unit 208 records “normal” in the status of the abnormality detection data of the flow line to be processed.

  On the other hand, in S1004, the detection unit 208 records “abnormal” in the status of the abnormality detection data of the flow line to be processed. In step S <b> 1005, the detection unit 208 calculates an abnormal action start point of a flow line to be processed, that is, a flow line (abnormal flow line) where an abnormal action is detected. Specifically, the detection unit 208 divides the flow line to be processed at the observation position at a time point that is a predetermined time back from the reference point, and performs abnormal behavior from the shaped flow line data of the divided flow line including the observation position of the reference point. It is determined whether or not. Note that the processing for determining whether or not the behavior is abnormal is the same as the processing in S1002. If no abnormal behavior is detected, the detection unit 208 moves the division position in the past direction, repeats the determination process, and calculates the observation time of the division position when the abnormal behavior is detected as the abnormal behavior start point. To do. Here, the abnormal action start point is a start point of the special action, and the process of S1005 is an example of the start point specifying process.

  For example, the flow line 502 in FIG. In this case, the detection unit 208 first determines whether or not the series “3, 4” closest to the reference point is abnormal behavior. If no abnormal behavior is detected, the detection unit 208 moves the division position of the series in the past direction, and determines whether the series “2, 3, 4” is an abnormal action. When the detection unit 208 determines that the series “2, 3, 4” is abnormal behavior, the observation unit 208 determines the observation time of the sampling observation point in the space 2 as the division position as the abnormal behavior start point.

  Next, in S1006, the detection unit 208 records the abnormal action start point and the abnormality detection point calculated in S1005 in the abnormality detection data of the flow line to be processed. Note that the detection unit 208 records the time at the time of processing in the abnormality detection point.

  FIG. 11 is a flowchart showing detailed processing in the display processing (S905) by the display processing unit 209. In step S <b> 1101, the display processing unit 209 acquires one shaped flow line data from among the shaped flow line data stored in the monitoring target flow line DB 205. In step S1102, the display processing unit 209 checks the flow line status acquired in step S1101. The status is a value updated in S1003 or S1004 of FIG. If the status is “normal” (normal in S1102), the display processing unit 209 advances the process to S1003. If the status is “abnormal” (abnormal in S1002), the display processing unit 209 advances the process to S1104.

  In S1103, the display processing unit 209 determines a time point a predetermined time (for example, 5 seconds) before the reference point as a display start point, generates display data for the flow line to be processed including the display start point, and then The process proceeds to S1106. Here, the fixed time is a predetermined length of time, and is set in the storage unit 104, for example. Note that the certain period of time may be a variable value depending on the density and speed of the moving body in the monitoring area. When setting a variable value, the information processing apparatus may shorten the display time as the density of the moving body is higher or the speed is higher. Thereby, when there are many flow lines, the overlap of flow lines can be reduced efficiently.

  On the other hand, in step S1004, the display processing unit 209 determines which of the abnormal action start point and the point of time before the reference point is a past time. If the time point before the predetermined time is a past time (a predetermined time before in S1004), the display processing unit 209 advances the processing to S1103. That is, in this case, the display processing unit 209 determines a point in time before a certain time as a display start point. If the abnormal action start point is a past time (the abnormal action start point in S1004), the display processing unit 209 advances the process to S1105.

  In step S1105, the display processing unit 209 determines the abnormal action start point as the display start point, generates display data for the flow line to be processed including the display start point, and then advances the process to step S1106. Thereby, it is possible to avoid that the length of the abnormal flow line becomes necessary and abnormally short. Note that the length from the display start point to the reference point is determined by determining the display start point in S1103 and S1105. That is, the processing of S1103 and S1105 is an example of processing for determining the length of a flow line as a movement trajectory including the start point, that is, an example of length determination processing for determining the length of the flow line that is a temporal length. is there. Here, the time length of the flow line determined in S1105, that is, the abnormal flow line is longer than the time length of the normal flow line. Furthermore, the abnormal flow line is a movement locus including an abnormal action start point. For this reason, it is possible to help a viewer such as a monitor grasp the progress and cause of abnormal behavior from the abnormal flow line.

  In addition, as another setting method of the display start point, the detection unit 208 determines the most past time point as the display start point within a range that is past the abnormal behavior start point and does not intersect with other flow lines. May be. Thereby, the maximum information regarding the abnormal flow line can be displayed within a range where the visibility of the flow line is not deteriorated.

  In step S1106, the display processing unit 209 determines whether or not the setting of the display start points for all the flow lines has been completed. If all the display start points have been set (Yes in S1106), the display processing unit 209 advances the process to S1107. If there is an unprocessed flow line (No in S1106), the display processing unit 209 advances the process to S1101 and repeats the process. In step S1107, the display processing unit 209 displays partial collinear lines from the display start point to the reference point on the flow line display screen for all flow lines. This is the end of the display process. Here, the process of S1107 is an example of a display process for displaying a flow line as a movement trajectory of a plurality of moving objects on the same screen for the length determined in S1103 or S1105.

  As described above, the information processing apparatus according to the first embodiment displays the length of the abnormal flow line longer than the length of the normal flow line on the flow line display screen. More specifically, the information processing apparatus displays a movement trajectory for a longer period than the fixed time for abnormal behaviors whose cause is difficult to grasp from the movement trajectory for a fixed time. That is, the information processing apparatus can provide information for inferring the progress and cause of abnormal behavior to a monitor or the like.

  As a first modification of the information processing apparatus according to the first embodiment, the information processing apparatus replaces the time from the reference point with the spatial length (distance) of the movement locus from the observation position at the reference point. Display start point may be determined on the basis of a typical length). In this case, in step S1103, the information processing apparatus determines, as a display start point, a past point for a certain movement distance of the moving object instead of a certain time. Then, the information processing apparatus displays a movement locus from the observation position at the display start point to the observation position at the reference point. That is, this process is an example of a length determination process that determines the length of the flow line that is the distance from the observation position at the display start point to the observation position at the reference point. In this case, the flow lengths of the normal flow lines are all the same length, and only the abnormal flow lines have different lengths. Accordingly, the information processing apparatus can display the abnormal flow line with more emphasis.

  Further, as a second modification example, in the present embodiment, the case of detecting an abnormal action (an ugly action) as an example of a special action has been described as an example, but the type of action to be detected is the embodiment. It is not limited to. The special action may be a normal action. When an action with a behavior characteristic that cannot be understood unless it is observed for a long time, the corresponding flow line is displayed longer than the other flow lines, and the progress of the action can be shown, contributing to quick grasp of the situation. be able to.

  As a third modification, the behavior model DB 207 may store a flow line pattern as a normal behavior model. In this case, the model construction unit 206 constructs a normal behavior model by, for example, extracting a frequent pattern from the flow line data set. As a frequent pattern extraction method, a known technique such as PrefixSpan can be used. Further, the transition of the existence space may be modeled stochastically using an HMM or the like. Further, in this case, in S1002, the detection unit 208 determines whether or not abnormal behavior has been detected by comparing the normal behavior model with the shaped flow line data. Specifically, the detection unit 208 determines that the subspace sequence of the full length of the flow line to be processed up to the reference point is included in the flow line pattern in the normal behavior model, and is not included. In the case, it is determined as abnormal. Also in this case, the detection unit 208 divides the flow line to be processed at the observation position at a time point that is a predetermined time back from the reference point, and determines whether the abnormal action is generated from the shaping flow line data of the divided flow line including the observation position of the reference point. Determine whether or not. In this case, it is assumed that the predetermined time is set to a value shorter than the length of time during which abnormal behavior can be detected.

  As another example, in the determination of abnormal behavior, the detection unit 208 may calculate the degree of detachment from the normal behavior model and determine that the abnormality is abnormal when the degree of detachment exceeds a certain level. For example, the detection unit 208 can use the sequence length of the portion that does not match the pattern in the normal behavior model in the shaping flow line data of the flow line to be processed as an outlier index. In this case, a slight deviation from normal due to individual differences is no longer detected as an abnormal behavior, so that the abnormal behavior can be detected with higher accuracy.

  Further, as a fourth modification example, the process in which the detection unit 208 detects abnormal behavior is not limited to the embodiment. As another example, the detection unit 208 replaces the flow line (movement trajectory) of the processing target, based on the movement history such as acceleration and motion vector of the person as the moving object corresponding to the processing target, You may determine the presence or absence of abnormal behavior.

  As a fifth modification, the information processing device is not limited to the abnormal behavior. The behavior of the detection target is not limited to the embodiment (abnormal behavior), as long as it is a behavior that is observed by an observer who observes the behavior (movement) of a person as a moving body. . As another example, it is assumed that the observer wants to observe a specific movement locus such as a movement locus of a child among the movement locus of many persons. In this case, a behavior model for detecting a child's movement trajectory is recorded in the behavior model DB 207 of the information processing apparatus. That is, the information processing apparatus is in a state in which the movement trajectory of the child to be observed is set as a special action. Thereby, the information processing apparatus can detect the movement history of the child set as the observation target as the special action.

  As a sixth modification, the display processing unit 209 has already determined the display start point when the status of the moving object is “abnormal” in step S1102 of the display process shown in FIG. The processing of S1103 to S1105 may not be performed. This process is an example of a moving body specifying process for excluding a moving body whose abnormal behavior has already been detected from the processing target of the flow line length determination process and the start point determination process.

(Second Embodiment)
Next, an information processing apparatus according to the second embodiment will be described. Hereinafter, the difference between the information processing apparatus according to the second embodiment and the information processing apparatus according to the first embodiment will be described. While the information processing apparatus according to the first embodiment displays normal flow lines and abnormal flow lines with different lengths of flow lines, the information processing apparatus according to the present embodiment further relates to abnormal behavior. By displaying the information, the information necessary for grasping the cause of the abnormal behavior by the monitor or the like is provided.

  FIG. 12 is a diagram illustrating the structure of the monitoring space. FIG. 13 is a diagram showing a display example of a flow line display screen corresponding to the monitoring space shown in FIG. FIG. 13 (A) to FIG. 13 (E) show the transition of the flow line in the flow line display screen over time. The monitoring space includes two straight passages as shown in FIG. It is assumed that the moving body can move in the upward direction and move in the right direction in the monitoring space, but it is prohibited to turn right at the backflow and the middle junction. As shown in FIGS. 13A to 13C, it is assumed that the moving body Q makes a right turn at a junction. In this case, as shown in FIG. 13C, the information processing apparatus according to the present embodiment displays the flow line 1301 of the moving body Q as a longer line than other normal flow lines. "Is also displayed as related information 1302 indicating the type of abnormal behavior.

  In FIG. 13C, a symbol 1303 indicates the position (reference position) of the moving body Q at the reference point. Note that the information processing apparatus changes the shape of the symbol 1303 indicating the current position of the moving body Q where the abnormal behavior is detected to a shape different from that of the moving body of the normal behavior, in order to indicate that the abnormal behavior is detected. . In the present embodiment, as shown in FIG. 13C, the information processing apparatus shows a round symbol for a moving body in which abnormal behavior is not detected, and a symbol for the mobile body in which abnormal behavior is detected. Shown in shape.

  FIG. 13D is a diagram showing a flow line display screen at a time 10 seconds after FIG. 13C. At this time, the abnormal action detection point 1311 is further displayed with a cross mark on the flow line display screen. The abnormal behavior detection point 1311 indicates the timing at which the abnormal behavior of the moving body Q is detected and the position thereof. For this reason, even when the position of the moving body Q is away from the abnormal action start point, the abnormal action section can be easily identified. Here, since an abnormality is detected at the time of FIG. 13C, a symbol indicating that is displayed at the position of the moving body Q in FIG. The information processing apparatus may further highlight the abnormal behavior section by changing the flow line display form before and after the abnormal behavior detection point.

  The related information 1302 is displayed between the abnormal action start point and the abnormal action detection point, and the display position does not change with time. This is because it becomes difficult to read information when the display position changes with the passage of time. When the related information is displayed at the fixed position in this way, the related information 1302 may overlap with other flow lines. In that case, the information processing apparatus hides the flow line overlapping the related information 1302, displays it as a dotted line, narrows the width, reduces the drawing density / lightness, and so on. Change the display mode to lower. Thereby, the visibility of the related information 1302 is maintained. In FIG. 13D, since the normal flow line 1312 overlaps the related information 1302, the normal flow line 1312 is changed to a dotted line display.

  However, when the flow line overlapping the related information 1302 is an abnormal flow line, it is not preferable to change the display form of the flow line as described above. Therefore, when the flow line overlapping the related information 1302 is an abnormal flow line, the information processing apparatus changes the display position of the related information 1302 to a position that does not overlap the abnormal flow line. FIG. 13E is a diagram illustrating a display example at a time after 10 seconds of FIG. Since the display position of the related information 1302 in FIG. 13D overlaps with the abnormal flow line 1321 in FIG. 13D, it can be seen that the display position of the related information 1302 is moved.

  FIG. 14 is an abnormal behavior model stored in the behavior model DB 207 according to the second embodiment. The abnormal behavior model includes a flow line pattern and an abnormal type. Then, the detection unit 208 compares the abnormal behavior model with the shaped flow line data, and detects the abnormal behavior when the flow line to be processed is included in the flow line pattern in the abnormal behavior model. FIG. 15 is a diagram showing symbol data. The symbol data is information that associates an action type, a symbol, and related information. The action type is roughly classified into “normal” and “abnormal”. Furthermore, the abnormal behavior includes a plurality of types of behavior such as “right turn violation” and “U-turn”, and there are behavior types corresponding to these. In the symbol and related information, a symbol displayed on the flow line display screen and related information are recorded. Note that the abnormal behavior model and the symbol data are preliminarily constructed and stored in the storage unit 104 or the like.

  FIG. 16 is a diagram illustrating an example of abnormality detection data according to the second embodiment. The abnormality detection data according to the present embodiment further includes an action type in addition to each data of the abnormality detection data according to the first embodiment. The action type is specified and recorded by the detection unit 208. FIG. 17 is a diagram illustrating an example of display data according to the second embodiment. The display data according to the present embodiment further includes an action type, a display form, and a related information display position in addition to each data of the display data according to the first embodiment. The action type is information specified by the detection unit 208. The display form is the density and line type of the line indicating the flow line, the related information display position is the position on the flow line display screen for displaying the related information, and both the display form and the related information display position are displayed. It is determined by the processing unit 209.

  A flow line display process performed by the information processing apparatus according to the second embodiment will be described. In S1002 (FIG. 10), the detection unit 208 of the information processing device specifies not only the presence / absence of abnormal behavior but also the behavior type of abnormal behavior when abnormal behavior is detected. In step S <b> 1004, the detection unit 208 records “abnormal” in the status of the abnormality detection data and also records the action type. Specifically, the detection unit 208 compares the partial space sequence of the flow line to be processed with the flow line pattern of the abnormal behavior model shown in FIG. Then, when the subspace sequence of the flow line to be processed is included in the flow line pattern of the abnormal behavior model, the detection unit 208 sets the abnormal type associated with the flow line pattern of the abnormal behavior model as the behavior type. Identify. On the other hand, when the subspace series of the flow line to be processed is not included in the flow pattern of the abnormal behavior model, the detection unit 208 determines that the abnormal type is unknown and sets “being unknown” as the behavior type of the abnormality detection data. Record.

  FIG. 18 is a flowchart showing detailed processing in the display processing (S905) by the display processing unit 209 of the information processing apparatus according to the present embodiment. The processes of S1801 to S1804 and S1806 are the same as the processes of S1101 to S1103, S1105, and S1106 in FIG. 11, respectively. In S1803 or S1804, the display processing unit 209 generates display data for the flow line to be processed including the display start point, and then advances the processing to S1805. In the processing of S1803 or S1804, the flow line ID, the display start point, the display flow line, the action type, and the display mode are recorded in the display data. In the display mode, “density 100, solid line” is recorded as an initial value.

  In step S1805, the display processing unit 209 records the action type specified by the detection unit 208 in the display data of the flow line to be processed, and advances the processing to step S1806. In S1806, when all the display start points have been set (Yes in S186), the detection unit 208 advances the process to S1807. In step S1807, the display processing unit 209 refers to the abnormal action start point and the abnormal action detection point, calculates the related information display position, and records this in the display data. For example, the display processing unit 209 may calculate the abnormal action start point or the abnormal action detection point itself as the related information display position, and as another example, calculates the intermediate position between both points as the related information display position. May be. Further, in the case of an abnormality that has been detected in the past, the display processing unit 209 may use the previous display position as it is as the related information display position without calculating the related information display position. Thereby, processing time can be shortened.

  Next, in step S1808, the display processing unit 209 refers to the display flow line data and the related information display position calculated in step S1807, and compares the related information display position and the abnormal flow line displayed on the flow line display screen. Check for any overlap. If there is an overlap (Yes in S1808), the display processing unit 209 advances the process to S1809. If there is no overlap (No in S1808), the display processing unit 209 advances the process to S1810. In step S1809, the display processing unit 209 changes the related information display position to a position where the related information and the abnormal flow line do not overlap, and also changes the related information display position of the display data.

  Next, in S1810, the display processing unit 209 refers to the display flow line data and the related information display position, and overlaps between the related information display position and the normal flow line displayed on the flow line display screen. Check for presence. If there is an overlap (Yes in S1810), the display processing unit 209 advances the process to S1811. If there is no overlap (No in S1811), the display processing unit 209 advances the process to S1812.

  In S1811, the display processing unit 209 changes the display mode of the normal flow line where the overlap occurs to a predetermined display mode. In the present embodiment, the display processing unit 209 changes the line density to a value that makes it thinner, and changes the line type from a solid line to a dotted line. Then, the display processing unit 209 changes the display mode of the display data according to the change content. It should be noted that a known spatial search technique can be used to determine the overlap between the relevant information and the flow line in S1808 and S1810.

  Next, in S1812, the display processing unit 209 displays all the flow lines stored in the monitoring target flow line DB 205 based on the display flow line data and the display mode. At this time, the display processing unit 209 also displays related information and symbols. The display processing unit 209 displays symbols in different shapes (display modes) for each action type (for each action type) based on the symbol data. Further, the display processing unit 209 may draw such that the abnormal flow line is in front when the abnormal flow line and the normal flow line overlap. This is the end of the display process. The remaining configuration and processing of the information processing apparatus according to the second embodiment are the same as the configuration and processing of the information processing apparatus according to the first embodiment.

  As described above, the information processing apparatus according to the second embodiment displays related information regarding abnormal flow lines. That is, information that can be used for grasping the cause of abnormal behavior can be provided.

  A modification example of the information processing apparatus according to the second embodiment will be described. The display processing unit 209 according to the embodiment indicates whether the behavior is abnormal or normal by the shape of the symbol shown together with the movement trajectory. However, the display process part 209 should just change the display mode of a movement locus | trajectory for every action classification, and the specific process for it is not limited to embodiment. As another example, the display processing unit 209 may change the line type of the movement locus for each action type. Further, the display processing unit 209 may change the size of the symbol for each action type. The symbol display position is not limited to the embodiment as long as it is a position where the correspondence with the movement trajectory is visible. As another example, the display processing unit 209 may display a symbol at the display start point.

(Third embodiment)
Next, an information processing apparatus according to the third embodiment will be described. Hereinafter, the difference between the information processing apparatus according to the third embodiment and the information processing apparatus according to another embodiment will be described. The information processing apparatus according to the third embodiment displays, as a movement trajectory, video data indicating the movement trajectory, that is, video data obtained by photographing the moving body, instead of the flow line display screen. By reproducing the video of the moving body, the supervisor can more easily understand what kind of abnormality has occurred.

  FIG. 19 is a diagram illustrating an example of a trajectory display screen in the present embodiment. The monitoring space and the moving body are in the same state as in FIG. 3D in the first embodiment. It is assumed that the moving body P has taken the same abnormal behavior as that shown in FIGS. 3 (B) to 3 (D). On the other hand, it is assumed that the mobile object R has taken normal action. The display processing unit 209 displays the symbols of the moving object P and the moving object Q in different shapes (display modes). Thereby, the supervisor can grasp | ascertain that the mobile body P took abnormal action. Furthermore, in the present embodiment, the information processing apparatus can display a monitoring video corresponding to a moving object designated by the user.

  FIGS. 19A and 19B are diagrams showing trajectory display screens 1900 and 1910 when the user designates the moving body R and the moving body P, respectively. In the monitoring video areas 1901 and 1911 on the right side of the trajectory display screens 1900 and 1910, the monitoring video images that capture the moving body R and the moving body P are reproduced and displayed as moving trajectories. The playback section bar 1920 indicates the playback section of the video, and the playback time is different between the video of the mobile body R and the mobile body P. For the moving body R, the latest video, that is, the video for the first time is reproduced, while for the moving body P, the video of a length that can grasp what abnormal behavior has been taken, that is, the first video. The video for the second time which is longer than the time is reproduced. The user can move the playback position of the video by moving the playback position pointer 1921.

  FIG. 20 is a flowchart illustrating detailed processing in the display processing (S904) by the display processing unit 209 of the information processing apparatus according to the third embodiment. In step S2001, the display processing unit 209 acquires the flow line of the moving object selected by the user. Hereinafter, the processing of S2002 to S2005 is the same as the processing of S1102 to S1105 of FIG.

  Next, in S2006, the display processing unit 209 reproduces the video of the moving object to be processed from the display start point. The display processing unit 209 refers to the observation data acquisition source illustrated in FIG. 4, acquires the corresponding video from the monitoring video DB 201, and reproduces and displays this. In addition, when a plurality of acquisition sources are recorded for one moving object, the display processing unit 209 continuously plays back images obtained by connecting a plurality of acquisition sources. The display processing unit 209 reproduces the video of the acquisition source in a section where there is only one acquisition source for the same observation point, and acquires a plurality of acquisitions in a section where a plurality of acquisition sources exist for the same observation point. Select one source video from the source and play it. At this time, the display processing unit 209 performs a contrivance such as preferentially reproducing the same acquisition source as the acquisition source in the previous section, or selecting and reproducing the acquisition source that reduces the number of video switching as a whole. Smooth video playback is possible. The remaining configuration and processing of the information processing apparatus according to the third embodiment are the same as the configuration and processing of the information processing apparatus according to other embodiments.

  As described above, the information processing apparatus according to the third embodiment can quickly reproduce the video of the section in which the abnormal behavior has occurred, and can more accurately grasp the cause of the abnormal behavior. A monitor or the like can determine whether the subject is a suspicious person or not knowing the target object by confirming the facial expression or clothes of the moving object on the video.

  As a first modification of the information processing apparatus according to the third embodiment, the information processing apparatus automatically reproduces the video of the mobile body that has taken an abnormal action instead of the video of the mobile body selected by the user. You may comprise. Thereby, the video of abnormal behavior can be reproduced without user operation, which is effective in an emergency. As a second modification, the information processing apparatus may display both the flow line and the video of the moving body. As a result, the supervisor or the like can first check the cause of the abnormal behavior by looking at the flow line, and further confirm the cause by reproducing the video.

(Other examples)
The present invention supplies a program that realizes one or more functions of the above-described embodiments to a system or apparatus via a network or a storage medium, and one or more processors in a computer of the system or apparatus read and execute the program This process can be realized. It can also be realized by a circuit (for example, ASIC) that realizes one or more functions.

  As described above, according to each of the above-described embodiments, it is possible to display the movement trajectory of the moving object so that the observer can easily distinguish the action to be observed from the other actions.

  The preferred embodiments of the present invention have been described in detail above, but the present invention is not limited to such specific embodiments, and various modifications can be made within the scope of the gist of the present invention described in the claims.・ Change is possible.

203 Data control unit 205 DB to be monitored
207 Behavior Model DB
208 Detection unit 209 Display processing unit

Claims (19)

Based on the moving history of the moving body, a detection unit that is set as an observation target and detects the special action of the moving body;
When the special action is not detected, the movement trajectory of the moving body is displayed with a predetermined first length. When the special action is detected, the movement trajectory is displayed as the first trajectory. An information processing apparatus comprising display processing means for displaying a second length that is longer than the length.   The information processing apparatus according to claim 1, wherein the first length and the second length are temporal lengths.   The information processing apparatus according to claim 1, wherein the first length and the second length are distance lengths. The detecting means detects a special action of each moving body when there are a plurality of moving bodies in the target space,
Of the plurality of moving bodies, the length of the moving locus of the moving body in which the special action is not detected is determined as the first length, and the length of the moving locus of the moving body in which the special action is detected. Further comprising length determining means for determining the second length as
4. The display processing unit according to any one of claims 1 to 3, wherein the display processing unit displays the movement trajectories of the plurality of moving bodies on the same screen by the lengths determined by the length determination unit. The information processing apparatus described in 1. In the case where the special action is detected, it further includes a start point specifying means for specifying a start point of the special action,
The length determining means determines a length including the start point as the second length,
5. The display processing unit according to claim 4, wherein, when the special action is detected, the display unit displays the movement trajectory with the second length determined by the length determination unit. Information processing device.   The information processing apparatus according to claim 4, wherein the special action is determined according to a type of the moving body. The detection means periodically detects the special action,
A mobile object for which a special action is newly detected by the detection means is specified as a processing target, and a mobile object specifying means for excluding a mobile object whose special action has already been detected from the processing target;
The information processing apparatus according to claim 6, wherein the length determining unit and the start point specifying unit perform processing on the moving object specified by the moving object specifying unit as a processing target. The detection means detects a plurality of types of special actions,
The information processing apparatus according to claim 1, wherein the display processing unit displays the movement trajectory in a display mode that is different for each type of special action.   The information processing apparatus according to claim 8, wherein the display processing unit displays type information indicating a type of the special action together with the movement trajectory of the moving object in which the special action is detected.   The information processing apparatus according to claim 9, wherein the display processing unit displays the type information at a position that does not overlap the movement locus of the moving body in which the special action is detected.   The display processing means displays the type information at a position overlapping the movement trajectory of the moving body in which the special action has not been detected, and the visibility of the movement trajectory overlapping the type information is reduced. The information processing apparatus according to claim 10, wherein the display mode of the movement locus is changed. Based on the moving history of the moving body, a detection unit that is set as an observation target and detects the special action of the moving body;
When the special action is not detected, an image showing the movement trajectory of the moving object is reproduced and displayed for a predetermined first time, and when the special action is detected, the video showing the movement trajectory Display processing means for reproducing and displaying for a second time longer than the first time. The information processing apparatus according to claim 12 , wherein the first time and the second time are values determined based on a distance of a movement trajectory. A starting point specifying means for specifying a starting point of the special action when the special action is detected;
Length determining means for determining a time including the start point as the second time;
13. The display processing means, when the special action is detected, displays an image showing the movement locus for the second time determined by the length determination means. Or the information processing apparatus of 13. The detection means periodically detects the special action, and further detects the special action of each moving body when there are a plurality of moving bodies in the target space,
A mobile object for which a special action is newly detected by the detection means is specified as a processing target, and a mobile object specifying means for excluding a mobile object whose special action has already been detected from the processing target;
The information processing apparatus according to claim 13, wherein the length determining unit and the start point specifying unit perform processing on the moving object specified by the moving object specifying unit as a processing target. An information processing method executed by an information processing apparatus,
A detection step for detecting a special action of the moving body set as an observation target based on a moving history of the moving body;
When the special action is not detected, the movement trajectory of the moving body is displayed with a predetermined first length. When the special action is detected, the movement trajectory is displayed as the first trajectory. And a display processing step of displaying a second length that is longer than the length. An information processing method executed by an information processing apparatus,
A detection step for detecting a special action of the moving body set as an observation target based on a moving history of the moving body;
When the special action is not detected, an image showing the movement trajectory of the moving object is reproduced and displayed for a predetermined first time, and when the special action is detected, the video showing the movement trajectory Including a display processing step of reproducing and displaying for a second time, which is longer than the first time. Computer
Detection means that is set as an observation target based on the movement history of the moving body, and that detects the special action of the moving body;
When the special action is not detected, the movement trajectory of the moving body is displayed with a predetermined first length. When the special action is detected, the movement trajectory is displayed as the first trajectory. A program for functioning as display processing means for displaying a second length which is longer than the length. Computer
Based on the moving history of the moving body, a detection unit that is set as an observation target and detects the special action of the moving body;
When the special action is not detected, an image showing the movement trajectory of the moving object is reproduced and displayed for a predetermined first time, and when the special action is detected, the video showing the movement trajectory For functioning as display processing means for reproducing and displaying for a second time that is longer than the first time.

Images